Chapter 5 - Cautionary Tales in CAD: When Tech Isn’t Enough

Not every story of a CAD adventure has a happy ending. Here are a few use cases that have some business lessons to teach us.

PTC's Mid-Market Misadventure: The Pro/JR Catastrophe

While Solid Edge and SolidWorks were successfully conquering the mid-market, the established high-end leader PTC was facing an uncomfortable reality: their customers were increasingly asking for more affordable alternatives to Pro/ENGINEER. PTC's response would become one of the industry's most cautionary tales.

In what can only be described as a catastrophic miscalculation, PTC launched Pro/JR in 1995—a stripped-down version of Pro/ENGINEER intended to compete with the emerging Windows-based solutions. The product was hampered by artificial limitations, poor performance, and a pricing strategy that satisfied neither high-end nor mid-market buyers.

Pro/JR's failure was so complete that it accelerated customers' migration to SolidWorks and other competitors. Rather than protecting PTC's market share, the initiative inadvertently validated the very products it was meant to compete against. The debacle reinforced PTC's eventual decision to focus exclusively on their high-end geometry engine.

In 2007, when PTC realized they needed a direct modeler for some scenarios to complement their parametric modeler, they acquired what was then CoCreate Solid Designer and rebranded it Creo Elements/Direct. This product, however, relies on a proprietary, ACIS-based kernel (see the timeline later in this article for details).

It wasn’t until 2019 that they purchased former PTCer and co-founder of SolidWorks, Jon Hirschtick’s Onshape (based on Parasolid kernel) for attacking the mid-market. I asked Steve Dertien, CTO of PTC, about whether Parasolid was still in use by Onshape:

Onshape as acquired is still based on Parasolid.  That's not easy to change, but we're also not exclusive.  We've already incorporated the Frustum kernel (acquired by PTC in 2018) for generative design as well as the Creo kernel for some other features. Similarly, Creo and all other CAD, do plug in other people's engines for features.  For example, we don't hide that we use Materialize (from Materialize NV in Belgium) in 3D Printing or ModuleWorks (from ModuleWorks GmbH in Germany)) for CAM simulation as well as Keyshot (from Luxion Inc in Costa Mesa, CA, USA) for rendering and Ansys for simulation.  Even when we added Ansys we still had to support the prior generations of simulations to maintain all the data feature compatibility. Every company decides where to build, buy and partner for technology in the stack where appropriate.

The Short, Somewhat Unhappy Life of CADDS5

CADDS5 was the final evolution of a CAD lineage dating back to CADDS1 in 1969, one of the earliest commercial drafting systems. Developed by Computervision, CADDS evolved through multiple generations — from 2D drafting to wireframe 3D (CADDS3) and eventually solid modeling (CADDS4X and CADDS5) in the 1980s. Unlike emerging kernels like Parasolid and ACIS, CADDS5 used a fully proprietary geometric modeling kernel, tightly integrated and never licensed or externalized. What made CADDS5 unique was its ability to support both direct modeling and parametric modeling, albeit via separate modules and executables — a powerful concept that prefigured later hybrid workflows.

After PTC acquired Computervision in 1998, it maintained CADDS5 for legacy industries like aerospace and shipbuilding, where long product lifecycles and regulatory lock-in made modernization difficult. But CADDS5 was eventually frozen at version 16.1 in 2013, with no future development. Its demise was due to several factors: a lack of modularity, no effort to license or replatform the kernel, and user resistance to its dated architecture and fractured modeling workflows. Meanwhile, the industry moved toward unified parametric-direct hybrid platforms like Solid Edge and, eventually, Creo.

Ironically, PTC’s later development of Creo did absorb some key lessons from CADDS5’s dual-mode modeling and large-assembly experience — but did so from a clean slate, not by reusing the CADDS kernel. The takeaway lesson is this: technological sophistication alone doesn’t ensure survival — adaptability, openness, and ecosystem strategy matter more than internal power. CADDS5 was ahead of its time in hybrid modeling but failed to evolve into an open platform others could build on. In the “kernel wars,” closed systems lost.

Forked at the Source: Autodesk’s Break from ACIS

When Autodesk set out to create Inventor—its answer to Pro/ENGINEER and SolidWorks—it knew it needed a robust 3D kernel. ACIS, then a rising player developed by Spatial Technology, was the obvious choice: proven, available, and already embedded in AutoCAD’s 3D extensions. But Autodesk made a bold move that would have long-term consequences: instead of fully committing to ACIS, they quietly forked the source code—creating their own derivative kernel called ShapeManager.

This gave Autodesk full control over the kernel’s evolution, independent of Spatial’s roadmap. But the story took a twist in late 2000, when Spatial was acquired by Dassault Systèmes, owner of SolidWorks—Autodesk’s rising nemesis. Suddenly, Autodesk found itself legally entangled with a competitor, accused of unpaid license fees on the forked code. Spatial sued. In 2003, Autodesk prevailed in court and retained royalty-free rights to its ShapeManager branch.

Meanwhile, Dassault cleaned up ACIS under Michael Payne’s leadership, fixing memory leaks and expanding functionality. But ACIS—once poised to challenge Parasolid—never regained the momentum it lost after both SolidWorks and Inventor abandoned it. While Spatial continues to license ACIS widely in mid-tier applications like BricsCAD and IronCAD, the kernel now sits behind the scenes, powering tools in markets where cost or compatibility matter more than cutting-edge modeling.

This episode isn’t just a legal footnote—it’s a striking example of kernel independence as strategic leverage. Autodesk’s decision to fork ACIS before Spatial’s acquisition gave it long-term autonomy, insulating Inventor’s roadmap from a now-rival platform. While rare, this approach has been mirrored by a few other vendors—notably CoCreate, whose SolidDesigner fork of ACIS still underpins Creo Elements/Direct. These cases serve as powerful reminders that owning your modeling core isn’t just a technical choice—it’s a business safeguard.

Today, ACIS continues to power some of the solutions of the CAD middle- and low-end markets through Spatial's OEM licensing program. Current ACIS-based applications include Dassault Systèmes' DraftSight, BricsCAD, and IronCAD (in this case they have a dual-kernel with ACIS and Parasolid) and various CAM and CMM software vendors.